As of 13 August 2026, AI can verify email addresses on your marketing list.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Export the marketing list as a CSV and keep the original file unchanged; include the email address, row reference and any business context needed to recognise duplicates.
   2. Remove unnecessary personal fields from the working copy, then upload the CSV to Apollo and use its available enrichment or email-quality workflow to obtain address-level results.
   3. Export the returned statuses and paste them, together with the original email column, into the prompt so the model can reconcile the results without losing the source rows.
   4. Ask the model to separate malformed, duplicate, role-based, disposable-looking, unverified and apparently usable addresses, with a reason for every suppression.
   5. Compare the model's cleaned addresses and suppression list against the original CSV and the statuses returned by Apollo, then manually inspect every row marked investigate or unverified.
   6. Save a new approved-send list containing only addresses that passed your chosen checks, and keep the original list, verification export and decision log together.
   7. Before sending, confirm separately that your campaign has the required consent or other lawful basis and that your suppression list is applied.

   Prompt

   I need to assess a marketing email list for sending in the UK. I will provide a CSV below.
   
   [PASTE CSV HERE]
   
   Do the following:
   1. Check each address for syntax errors, obvious typos, duplicate entries, disposable-looking domains, role addresses such as info@ or sales@, and missing values.
   2. Do not claim that an address is deliverable, that a mailbox exists, or that consent exists unless I provide a result from a live verification service or evidence for that claim.
   3. If you have no live DNS or email-verification access, label those checks as unverified rather than guessing.
   4. Return a table with the original address, a cleaned address where appropriate, status, reason, and recommended action: keep for further checking, suppress, or investigate.
   5. Preserve the original address and row reference so I can reconcile your output with the source list.
   6. Do not alter names, companies or other fields unless you show the original and changed value separately.
   7. Summarise the number of rows in each status only from the supplied data, and list every address that needs manual review.
   8. State clearly that this checks data quality and available verification evidence, not lawful basis, consent, or guaranteed future deliverability.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.